🔥 Wildfire Watch

Open-source wildfire detection and suppression system for edge deployment

Overview

Wildfire Watch is an automated fire detection and suppression platform that runs entirely on your local network. It uses AI-powered camera monitoring to detect fires and automatically activates sprinkler systems when multiple cameras confirm a fire.

Key Features

• 🎯 Multi-camera consensus - Prevents false alarms
• 🚀 Edge AI acceleration - Supports Coral, Hailo, NVIDIA GPUs
• 💧 Automated pump control - GPIO-based sprinkler activation
• 📹 24/7 recording - Frigate NVR with motion detection
• 🔒 Secure by default - TLS encryption (requires certificate setup)
• 🔄 Self-healing - Automatic camera discovery and failover

Quick Start

1. Clone and Setup

# Clone repository
git clone https://github.com/your-org/wildfire-watch.git
cd wildfire-watch

# Copy environment template
cp .env.example .env
# Edit .env to configure your settings

2. Security Setup (Choose One)

Option A: Development/Testing (Insecure)

# Use default certificates - INSECURE, for testing only!
# No action needed, default certs are included

Option B: Production (Secure)

# Generate secure certificates and enable TLS
./scripts/configure_security.sh enable
# This will guide you through certificate generation if needed

# Or manually:
./scripts/generate_certs.sh custom
echo "MQTT_TLS=true" >> .env

3. Deploy Services

# Deploy with Docker Compose
docker-compose up -d

# Verify all services are running
docker ps

# Check service logs
docker-compose logs -f

4. Access Services

• Frigate NVR: http://your-device:5000
  • Default credentials are shown in logs: docker logs security_nvr | grep Password
• MQTT Broker: Port 1883 (insecure) or 8883 (TLS)

⚠️ Security Warning: Default certificates are INSECURE. Always generate custom certificates for production use. See Security Guide.

5. Verify Security Configuration

# Check current security status
./scripts/configure_security.sh status

# Example output:
# ✓ TLS is ENABLED
# ✓ Using CUSTOM certificates
# ✓ MQTT broker is using TLS port 8883

Documentation

Getting Started

• Raspberry Pi 5 Guide - Recommended for most users
• Linux PC Guide - For x86 systems with GPU
• Hardware Requirements - Complete component list

Service Documentation

• Camera Detector - Automatic camera discovery
• Security NVR - Frigate configuration
• Fire Consensus - Multi-camera validation
• GPIO Trigger - Pump control system
• MQTT Broker - Communication hub
• Camera Telemetry - Health monitoring and metrics

Advanced Topics

• Model Converter - Custom AI models
• Multi-Node Setup - Scale to large properties
• Troubleshooting - Common issues
• Coral TPU Python 3.8 - Important compatibility info

System Architecture

┌─────────────┐     ┌─────────────┐     ┌─────────────┐
│  IP Cameras │────▶│   Camera    │────▶│  Frigate    │
│   (RTSP)    │     │  Detector   │     │    NVR      │
└─────────────┘     └─────────────┘     └─────────────┘
                            │                    │
                            ▼                    ▼
                    ┌─────────────┐     ┌─────────────┐
                    │    MQTT     │◀────│    Fire     │
                    │   Broker    │     │  Consensus  │
                    └─────────────┘     └─────────────┘
                            │                    │
                            ▼                    ▼
                    ┌─────────────┐     ┌─────────────┐
                    │    GPIO     │────▶│    Pump     │
                    │  Trigger    │     │   System    │
                    └─────────────┘     └─────────────┘

Service Architecture (Refactored)

All services now use standardized base classes for improved reliability:

• MQTTService: Automatic reconnection with exponential backoff
• HealthReporter: Standardized health monitoring and metrics
• ThreadSafeService: Safe thread management and shutdown
• ConfigBase: Centralized configuration management

Configuration

Essential Settings (.env)

# Camera credentials (comma-separated username:password pairs)
CAMERA_CREDENTIALS=username:password,username2:password2

# Fire detection
CONSENSUS_THRESHOLD=2      # Cameras required for consensus
MIN_CONFIDENCE=0.7         # AI confidence threshold

# Pump control
MAX_ENGINE_RUNTIME=1800    # 30 min (adjust for your water tank size!)
REFILL_MULTIPLIER=40       # Refill duration multiplier

# Hardware
FRIGATE_DETECTOR=auto      # auto|coral|hailo|gpu|cpu

# Service health monitoring
HEALTH_REPORT_INTERVAL=60  # Health report frequency (seconds)

# MQTT reconnection (all services)
MQTT_RECONNECT_MIN_DELAY=1.0   # Min reconnection delay
MQTT_RECONNECT_MAX_DELAY=60.0  # Max reconnection delay

See Configuration Guide for all options.

Deployment Options

Docker Compose (Recommended)

docker-compose up -d

Balena Cloud

balena push wildfire-watch

Kubernetes

kubectl apply -k k8s/

Hardware Support

Accelerator	Performance	Power	Recommended For
Coral TPU*	15-20ms	2W	Low power, always-on
Hailo-8L	20-25ms	2.5W	Raspberry Pi 5
Hailo-8	10-15ms	5W	High accuracy
NVIDIA GPU	8-12ms	15W+	Multiple cameras
CPU Only	200-300ms	5W	Testing only

*Note: Coral TPU requires Python 3.8 for tflite_runtime compatibility

Contributing

See CONTRIBUTING.md for development setup.

Support

• 📖 Documentation
• 💬 Discussions
• 🐛 Issues

License

MIT License - See LICENSE for details.

Disclaimer

This system is provided as-is for educational and experimental use. It does not guarantee fire prevention or property protection. Always follow local fire safety regulations and consult professionals for critical safety systems.